Digital Signal Processing Reference
In-Depth Information
Fig. 17.4. Waveform of a digital
audio signal stored in the
testaudio1.wav file.
1
0.5
0
−0.5
au
di
o
k
−1
0
5000
10 000
15 000
20 000
25 000
in the propagation of pressure waves. The human auditory system processes
the air waves and uses the information contained in the pressure variations to
extract audio information from the wave. It is possible to process sound waves
directly, as in a microphone, which converts sound to electrical signals that
are amplified and played back using a loudspeaker. The term audio refers to
electronically recorded or reproduced sound, while digital audio is obtained by
the sampling and quantization of an analog audio signal. The waveform of an
audio signal is shown in Fig. 17.4.
An audio signal is described using two properties. The first property is pitch,
which describes the shrillness of sound. Pitch is directly related to the fre-
quency of the audio signal and the two terms are used interchangeably. The
second property is the loudness, which measures the amplitude or intensity
of the audio signal using the decibel (dB) scale. Generally, the audible inten-
sity of an audio signal varies between 0 and 140 dB, where 0 dB represents
the lower threshold of hearing, below which a human auditory system is inca-
pable of hearing any sound. Typical office environments have an ambient audio
level of about 70 dB. Audio above 120 dB is very loud and is injurious to
humans.
Sound generated from physical phenomena contains frequency in the range
0-10 GHz. Since the human auditory system is only intelligible to sound fre-
quencies between 20 Hz and 20 kHz, most audio signals record sound within
this audible range and neglect any higher-frequency components. For example,
the digital audio stored on an audio compact disc is obtained by filtering the
CT audio by a lowpass filter with a cut-off frequency of 20 kHz, and the fil-
tered signal is sampled using a sampling rate of 44.1 ksamples/s. The number
of quantization levels used to produce digital audio depends upon the appli-
cation and may vary from 4096 levels obtained with a 12-bit quantizer, to 65
536 levels with a 16-bit quantizer, to 4 million levels with a 24-bit quantizer.
Higher numbers of quantization levels result in lower distortion and more pre-
cise reproduction of the original sound.
17.2.2 Formats for storing digital audio
Digital audio is available in a wide variety of formats, such as the au, wav, and
mp3 formats. Both au and wav formats store audio in the uncompressed form,
while mp3 compresses audio using Layer 3 of the MPEG-1 audio compression
Search WWH ::
Custom Search